Modification of GFAP-isoforms expression in human and mouse
astrocytes.
Student: Andreas Kompatscher (5600626)
akompats@gmail.com
Date: 11-08-2011
Supervisor: Regina Kanski
r.kanski@nin.knaw.nl Introduction
In recent years the subventricular zone (SVZ) has been identified as being one of the two areas in the mammalian brain, where adult neurogenesis takes place (Bovetti et al 2011). Neurogenesis in the SVZ is thought to be primarily a process concerned with supplying new neurons to the olfactory bulb (Bovetti et al. 2011). These are formed by adult neural stem cells in the SVZ and will then migrate to the OB to be integrated in that network. The pluripotent neural stem cells where identified as being astrocytes because, they express a marker for astrocytes, the glial fibrilliary acidic protein (GFAP). They reside in the SVZ and are located along the length of the lateral ventricles. This protein has multiple splice forms. One of these, namely GFAP-delta, is exclusively expressed in neurogenic astrocytes in the SVZ (Roelofs et al. 2005). Which suggests that astrocytes that express this specific isoform are the neural stem cells of the adult human brain (Middeldorp et al. 2010).
GFAP is an intermediate filament (IF) protein, which is a component of the cytoskeleton in astrocytes. IFs cellular assembly and organization provide cells with a mechanism to handle mechanical stress and deformation, and most importantly facilitates the motile activities of the cytoskeleton (Helfand et al. 2003; Roelofs et al. 2005). IF complex formation usually involves more than one IF protein. It is highly dependent of cell type, which IF proteins are expressed to form the network. In astrocytes, the pre-dominant IF proteins that make up the IF complex are GFAP, nestin, vimentin and synemin. The exclusive expression of GFAP in astrocytes makes this protein an excellent marker for these cells. Here we will focus on the different functions of GFAP, particularly in mouse studies. In mice, it was found that GFAP mRNA and protein expression are increased in astrocytes in to different types of injuries. This is a process known as astrogliosis. Simultaneously, the astrocytes undergo hypotrophy of their cellular processes. In a mouse study, it was found that GFAP and vimentin are of fundamental importance for the maintenance of CNS homeostasis (Pekny, 2001; Preng et al. 2008). Specifically GFAP is involved in astrocyte volume regulation, glial scar formation and anchoring glutamate transporters to the plasma membrane to facilitate neurotransmitter recycling. Also, mice that have a GFAP deficiency exhibit improved posttraumatic regeneration of neuronal synapses (Wilhelmsson et al. 2004; Preng et al. 2008). Mutations in the GFAP gene are the main cause of Alexander disease, where cytoskeletal defects lead to widespread astrocytic dysfunction. Furthermore it was found that in mice, astrocytes in the subventricular zone have the ability to become neurogenic astrocytes, which means they are pluripotent and have stem cell properties.
It is known that in humans, GFAP has multiple splice variants. The canonical isoform is GFAP-alpha which has nine exons. One of the alternative isoforms is GFAP-delta and is found exclusively in
neurogenic astrocytes (Roelofs et al. 2005). It differs from GFAP-alpha, since it does not express exons 8 and 9 and has a modified 7th exon (exon 7+) (Middeldorp & Hol, 2011). The modification
substitutes the 42 amino acid long c-terminal protein sequence of GFAP-alpha with a unique 41 amino acid sequence. This sequence is conserved across mice, rats and humans. GFAP-delta is a splice form that negatively influences the IF network, because of its changed structure. GFAP-delta mostly stays in a soluble form and when it is incorporated into the IF network it disrupts the
formation of a functioning network. It is hypothesized that this network has different properties than a network without GFAP-delta. Since these cells are neuronal stem cells and migrate through the brain it is thought that changes in the ratio of GFAP-alpha and delta can affect cell motility. As the network becomes less rigid it could allow the cell to be more mobile. The expression of GFAP-isoforms such as GFAP-delta can be manipulated using RNA interference (RNAi).
RNAi is a tool that allows for a sustained knockdown of a specific protein. It functions through small RNA fragments that bind specifically to the desired target mRNA, which disrupt the transcription of the protein. At first, candidate oligonucleotides, that encode for shRNAs, are designed and tested by cloning them into a vector, which is than used to transfect or transduce the desired cell type. Inside the nucleus the shRNA will be transcribed from the vector and gets transported to the cytoplasm through the nuclear pore by exportin 5. In the cytoplasm the shRNA is then cut by dicer. The double stranded shRNA gets dissociated and only the antisense strand gets loaded into a RNA inducing Silencing Complex (RISC). If an mRNA is complementary to the target sequence of shRNA the RISC complex will start to cleave the mRNA, preventing any translation to protein. The cleaved mRNA is then degraded by cellular RNAses.
To knock-down GFAP-isoforms in human and mouse cells the vector needs to be introduced to the cell and get incorporated into the nucleus. There are various methods that one can use. Two
techniques that are used in this project are the standard lipofectamine transfection and the lentiviral transduction of cells. A lipofectamin transfection is a useful tool to quickly obtain results for the knock-down capabilities of candidate vectors. Lentivirus is used to greatly enhance transduction efficiency compared to the relatively low transfection succes with standard lipofectamine transfections. The specific use of lentiviruses instead of other viral carriers as for example adeno associated virus has multiple reasons. An important reason is that AAV (adeno associated virus) does not integrate the DNA of interest into the host genome, whereas lentivirus (LV) does (Manjunath et al. 2009). For in vivo experiments this means that using LV ensures that the shRNAs are longterm expressed. Other advantages of using LVs is the ability to transduce any cell not only cells that are proliferating and being able to put in a larger vector compared to AAV. Since lentiviral transduction is essential for our experimental design, a short overview of lentiviral production and functioning is given.
A lentiviral particle is comprised of multiple elements, most of which are responsible for protecting the DNA or RNA. LV particles are enveloped in a lipid membrane. Embedded in that membrane are the envelope proteins, which interact with specific receptors on the surface of a target cell. The envelope used for the LV used in this experiment is not the natural envelope protein but the vesicular stomatitis glycoprotein (VSV-G). This protein enhances the tropism of the virus, increases the stability of the particle ensuring that it can withstand ultracentrifugation. This is necessary for virus isolation and makes the particle more resistant to freeze-thawing cycles. The most abundant protein in the virus particle is GAG, and it cleaved into three different structural proteins that form
layers underneath the lipid membrane. The outer layer is the matrix that surrounds the viral core. The core itself is delimited by a protein shell composed of capsid proteins, which enclose the nucleoprotein complex. The viral nucleus contains the vector with the shRNA targeting GFAP-isoforms together with three enzymes encoded by the POL genes. These are reverse transcriptase, integrase and protease (de las Mercedes Segura et al. 2006). Another important component is the REV gene which is important for the posttranscriptional protein regulation.
To produce virus, multiple vectors are transfected into HEK 293 T cells that contain the GAG/POL , REV and VSV-G genes for the structure, enzymes and envelope of the LV particle. To enhance biosafety researchers started to delete many non-essential genes from the HIV-1 virus making it less capable of recombining into a virus that had the ability to further replicate in the host cell. In the same vein the virus production vector with the GAG, POL, REV and envelope genes was split into two separate vectors. One vector contains the GAG/POL and REV genes, while the other has the
envelope. Several recombinations of the viral genes would be necessary to produce replication competent virus. To further increase safety the Rev genes were cloned into a separate vector. The goal of this project is to clone GFAP-isoform specific vectors that are capable of knocking-down GFAP-isoforms GFAP-alpha and delta in primary astrocytes and can be delivered to these cells by viral transduction in vitro. Furthermore the knock-down capabilities of general shRNA candidates, which target canonincal exons of the GFAP gene, provided by the RNAi consortium (a library of shRNA vectors) will be tested and virus will be made of the vectors that are efficient in knocking down GFAP. These vectors can then be used as a control in experiments involving the isoform-specific constructs that are used to investigate the function of GFAP-alpha and delta and can be used in experiments that investigate the function of GFAP in general.
Materials and methods
Cloning; The triple ligation approach
Candidate vectors were made through cloning to acquire GFAP-delta specific vectors. The first approach uses the lentiviral vector (pRRLsinPPT) lv-GFP-ictr as a backbone. The control shRNA (ictr) and promoter were removed and isolated from the rest of the backbone. The backbone was cut with restriction enzymes XMA I & CLA I. To ensure that all vectors are cut the incubation step was 3 hours for each enzyme, diminishing the amount of colonies on the control plate significantly. The lv-GFP backbone was then extracted from a gel and ready for use. The H1 promoter is used to promote transcription of the shRNA in eukaryotic cell lines. It was extracted from a different vector (pSuper) and cut out with restriction enzymes XMAI and BGLII. Candidate shRNAs were synthezised by Eurogentech and contain the restriction sites for CLAI and BGLII. The lv-GFP backbone, H1 promoter and shRNA candidate were ligated together using a triple ligation (1:5:5). The vectors were all transformed into E. Coli. strains GT116 and Stabl2 (Invitrogen) with a heat shock protocol. The heat shock was done at 42 ⁰C followed by a 2 minute incubation on ice. After the heat shock, the transformed bacteria are incubated for an hour at 37 ⁰C for the GT116 bacteria and 1,5 hours at 30 ⁰C for Stbl2 in 900 μl of LB medium and then plated on LB-agar (ampicillin) plates. These were then incubated overnight at 37 ⁰C (GT116) or 30 ⁰C (Stbl2). The colonies are picked on the next day to grow cultures.
The vectors are then isolated from the bacteria using a mini-preparation kit (Qiagen). To find out which colonies contain a correctly assembled vector the isolated vectors are digested with restriction enzymes BGLII and EARI. Digesting the vectors with these enzymes gives a different band pattern on an agarose gel compared to the pattern of the original vector that is digested with the same
enzymes. The possible candidate vectors that show the expected band pattern are amplified with PCR and send for sequencing. We used a Big Dye Terminator PCR protocol with primers targeting the H1 promoter. The sequencing is done to determine whether the vector is correctly assembled and contains the correct shRNA sequence. Because these vectors contain oligos that have a part encoding for the hairpin structure in a shRNA, it is difficult to sequence these. A betaine protocol was used to enhance the sequencing efficiency.
Cloning; The subcloning approach
Because this initial approach yielded no usable candidates for the lv-GFP-delta (human) candidates, a different approach was chosen. The new approach consists of using the PLKO1 vector that was provided by the RNAi consortium as a backbone and clone GFP and the GFAP-delta oligonucleotide into that vector. First the GFP is cut out of a vector (pcDNA3GFP) using restriction enzymes ApaI and BamHI and is then subcloned into pBluescript. This vector is used for subcloning before the GFP is cloned into the PLKO1 lentivector. GFP is cut out of pBluescript with restriction enzymes BamHI and KpnI and ligated into the PLKO1 lentivector. After cloning in GFP, the GFAP-delta shRNAs will be cloned into the lentivector using a similar protocol. The bacteria used were E. Coli strain DH5α instead of stbl2 because these bacteria are more competent.
Transfecting with general knock-down GFAP candidates
The general GFAP knockdown vectors were first tested using a lipofectamine transfection. On the first day, DBT cells (Mouse astrocytomas; Sexton et al. 2009) or U343 (human astrocytomas; Dirks et al. 1997) were plated in a 24 well plate. Each well was plated with 80.000 cells. On the second day, the transfection was done using 0,64 ug of DNA per well together with 1,6 ul of lipofectamine. In total five candidate vectors and five controls were used in the transfection. The controls were a vector containing a GFP control, one containing a non-targeting sequence, a vector containing cherry and untransfected cells. After 36 hours, the RNA is isolated from the cells using a Trizol based RNA isolation protocol. The RNA is transcribed into cDNA with a Quantitect reverse transcriptase kit and the cDNA is then used for Q-PCR. All Q-PCRs are done with fluorescent sybergreen dye and the primers used were targeting GFAP. These primers, that are specific for GFAP in general, target a sequence in exon 6, which is an exon that is shared by nearly all GFAP-isoforms. Primers that are specific for GFAP-delta were also used. These target the transition point between exon 7 and exon 7+. Actin primers were used as a control and as a means of quantifying the amount of GFAP.
Transduction Second Generation
Transient transfections alone are not enough to determine whether gene expression has been knocked-down. However these transient transfections do not allow for investigation of protein knock-down. To remedy this, it is advantages to transduce cells with lentivirus, ensuring that more then 90% of all cells are infected with a vector containing the shRNA. For the GFAP-isoform specific
lentiviral vectors, second generation virus production will be used. On day one 10 million HEK 293T cells are plated on 15 cm dishes. These cells were between 85% and 95% confluent and are
incubated for 24 hours. On day two, two hours prior to transfection the medium is changed. The transfection mixes contain the envelope (Rev), packaging (gag/pol) and the delta shRNA vector, which is dissolved in NaCl 0,9%. Pei dissolved in NaCl 0,9% is added to the vector mix, which is then incubated for 15 minutes. After the incubation, 2 ml of the mix is evenly distributed on each plate. After 16 hours the medium is changed and the fluorescence should be between 80%-90%, giving an indication how successful the transfection was. On the fourth day, the virus is isolated. This is done by collecting the medium from the cells, which now contains the produced virus. The medium containing the virus is then spun down at low speed (500 g for 5 minutes). The cleared supernatant is filtered through a 0,22 um filter directly into a Beckmann tube. The medium with virus is then spun for 2.30 hours at 20.000 rpm. After centrifugation the supernatant is poured off and the pellets are allowed to dry for 10 minutes. PBS with 0,5% BSA is than added to the pellet and left to incubate for 10-15 minutes. After the incubation, the pellet is gently resuspended in the added fluid. The BSA is added to prevent the virus from sticking too much to the surface of the storage tubes. Using BSA is only recommended when the virus is supposed to be used in vitro. The pellet is then gently resuspended a second time and aliquoted into screw-cap eppendorfs and frozen at -80º.
Virus Titering
To quantify the amount of virus particles that are capable of transducing cells, it is possible to titer them. First HEK 293T cells (100.000 cells/well) are plated on a 24 well plate. 24 hours after seeding the cells are transduced. A 1:10 dilution series is set up, which has a range from 1*10-4 to 1*10-7. The
different virus dilutions are used to transduce the cells. After 16 hours the medium is changed and at this time point around 50%-60% of the cells of the 1*10-4 dilution should be fluorescent. 24 hours
later, the medium is removed and the cells are fixed with PFA (incubation time 15 minutes). The PFA is then removed and the wells are washed twice with PBS. The plate was than examined under a fluorescence microscope (Axiovert 200M) to assess the transduction capabilities of the newly produced virus. The dilutions 10-6 and 10-7 are used to assess this. Five pictures of different locations
in the well are made of each dilution. The pictures were taken at a 10X magnification with an exposure time of 3000ms. For every picture, the amount of transduced and therefore fluorescent cells are counted and an average is calculated. To calculate the total amount of transduced cells one has to calculate the surface area of the pictures that were taken, which should than be divided by the surface of one well (190mm2), resulting in the amount of fields per well. The titer is than the ‘average
amount of GFP positive cells’ * ‘the amount of fields/per well * 2, which results in the total amount of transduction capable units per ml. The titers for the two dilutions are than compared and should not differ from each other except for one order of magnitude, which stands for the difference between the dilutions. The average of the two dilutions is then taken.
Third Generation
The general GFAP knock-down vectors (PLKO1-GFAPshRNA) are optimized for third generation virus production. Third generation virus production uses three virus vectors opposed to the two used in second generation virus production. This protocol requires lipofectamine instead of PEI as its lipid-carrier. Since lipofectamine is more expensive compared to the relatively cheap PEI, we first applied our third generation protocol with the three third generation vectors, but using PEI instead of
lipofectamine. Regrettably the resulting viral titers were very low (> 106) and thus yielded not enough units capable of
transduction. Lipofectamine transfections were than started on a small scale (1,8 million cells on a 6 cm dish) following the third generation protocol. This also resulted in low titers for the third generation plasmids. A different approach was then taken, by using the second generation protocol along with PEI as the lipid-carrier and the second generation
packaging and envelope vectors. This will enhance the chance that all vectors are transfected into a cell and was indeed successful.
Results
Cloning of GFAP isoform specific vectors
The cloning of a GFAP-delta specific construct has as of yet not resulted in any viable candidates. Multiple approaches have been tried but as of yet no actual GFAP-delta specific constructs were made. The original triple ligation approach was a strategy that did not result in any viable candidates for the human shRNAs (Fig. 1). Progress was made with the mouse GFAP-delta vectors and possible candidates were produced. However due to recombination these candidates lost their fluorescence. To resolve the problems with the triple ligation, we switched to a new approach and different bacteria (see Materials and Methods ‘subcloning approach’). Further
details concerning the various problems and solutions that are involved with the different cloning approaches can be found in the discussion section.
Characterization of the GFAP shRNA targets
For the general GFAP knock-down project we received five candidate vectors for both mouse and human. Vectors encoding the shRNA sequences where provided by the RNAi consortium (Fig. 2). However, which GFAP exon was targeted by the shRNAs was not known. Therefore, the target sequences of the shRNAs, were aligned to the GFAP sequence (NCBI Blast). Figure 3 shows that the candidates for mouse are targeting exons, between exon 1 and 6 that are shared by all isoforms. 90603, 90604, 90605, 90606, 90607, target in order GFAP exons 4, 3, 4, 1, 6. For the human candidate vectors D2 and D3 also target exons that are shared across isoforms. D2 and D3 target exons 4, and 2. However, vectors D1, D4 and C12 target exons 8 and 9 (Fig. 3), which are not shared by GFAP-delta. Therefore these two vectors are possible candidates for a specific knock-down of GFAP-alpha.
6 90603 90604 90605 90606 90607 D1 D2 D3 D4 C12
Validation of GFAP knock-down
We tested the knock-down capabilities of the general GFAP knock-down vectors to assess, which of the five candidates, for both mouse & human, showed efficient downregulation of GFAP mRNA. Figure 4 shows that candidates 90603, 90604 and 90606 resulted in a knock-down of murine GFAP in DBT cells compared to a GFP control.
Candidate-vector 90607 and 90605 are discarded, since 90607 does not give significantly lower GFAP levels compared to the control or other candidates. 90605 was discarded since the vector targets the same exon as 90603 and it is preferable to continue with candidates that each target a different exon. This ensures that a knock-down is caused by targeting the gene and not because targeting the exon activates another pathway that causes the knock-down. For the human GFAP knock-down candidates, virus production will be started with D3, D4 and C12 (Fig. 5). D4 and C12 were specifically chosen, because these vectors target exon 9.
This means that these vectors do not target GFAP-delta and are therefore GFAP-alpha specific. This is shown in Figure 6. Construct D3 gives knock-down of GFAP-delta in two out of three experiments whereas; D1, D4 and C12 seem to slightly up regulate GFAP-delta. D2 was discarded, since it up
regulates GFAP- delta and does not knock-down GFAP-alpha. However, the strong upregulation of delta by this shRNA is interesting, since it changes the ratio between alpha and delta in the favor of delta. This is interesting for experiments that want to induce GFAP-delta in for example neurogenic astrocytes. D1 will not be used since it can target other proteins then GFAP. In figure 8 it is indicated, that multiple genes are targeted by GFAP knock-down candidate D1. Many of these genes are involved in cell signaling and, if targeted by a shRNA, could change cell properties. Even though the match with the genes it targets is only 86% in most cases it poses to great a risk for off target effects. Therefore, D1 will be discarded in favor of alpha specific candidates D4 and C12. Figure 7 shows the data as a ratio of alpha and delta and it confirms that alpha is knocked-down more compared to delta, which results in a decrease of the alpha versus delta ratio. Transfection of D2 also results in a lower alpha to delta ratio. In this case, as shown in figure 5 and 6 delta is upregulated and alpha is not knocked-down, which results in the low ratio score.
Results for the virus production
At first, a third generation virus production protocol was used with the third generation vectors provided by the RNAi consortium. These experiments were done on a smaller scale than the second
generation virus production, since the expensive lipofectamine had to be used. However producing virus with this protocol was not successful and possible explanations will be given in the discussion section. Producing virus with the second generation protocol was successful. However, titers are lower when compared to second generation vectors.
Table 2 shows the titer of the original lentivector containing mCherry and the non-coding control shRNA (lv-cherry-ictr). This is a second generation lentivector and virus was produced with a second generation protocol. This is used as a measure to compare the titers of the third generation
lentivector. The titer for PLKO1-tGFP (Table 2) shows that indeed titers are 10 times lower for the third generation vectors compared to second generation vectors.
Discussion
Cloning GFAP-delta specific constructs for Mouse and Human
The first approach using the lentivector pRRLsinPPT to clone in the H1 promoter and the oligo with a triple ligation protocol proved to be problematic. Using E. Coli strain GT116 resulted in very few colonies and after isolating the vector and analysis on a gel, we could conclude that these colonies where backligations of the cut lentivector. Possible explanations are that triple ligations have a low succes rate of being assembled correctly and also have a high frequency of back ligations. This was especially evident in the mouse GFAP-delta clones. Furthermore it is known that lentiviral vectors have a tendency to recombine. A problem that hampered vector production.
This plasmid instability stems from deletions of regions between the long-term repeats (LTR) of a vector. Caused presumably by homologous recombination events. Such deletions remove most of the important lentiviral sequences and results in a smaller plasmid, which usually only contains the antibiotic resistance and replication origin. These recombinant plasmids are strongly selected, and recovery of the desired clones is difficult (Chakiath & Esposito, 2007).
To remedy this problem, we used a different strain named Stbl 2, which is specifically designed to be more stable when cloning with vectors that are known for being problematic like lentivectors. Stbl 2 enhanced stability is therefore used to prevent recombination in these lentivectors. However these strains are often slow growing, sensitive to phage T1 and difficult to make competent. Producing competent Stbl 2 was indeed problematic and was done successfully after switching to a protocol that uses TSS buffer. Although a control plasmid does give good transformation efficiency the
transformation of ligation mixes has as of yet not yielded any colonies. A switch to more easy to grow bacteria (DH5α) was made. Since the PLKO1 vector is used, it is expected that the recombination problem can be avoided as this contruct is more stable than the pRRLsinPPT vector. An other possible solution is to use E. Coli strain MDS42. This strain has a reduced genome and is specifically designed to deal with recombination problems and low efficiency (Chakiath & Esposito, 2007).
Testing the knock-down capabilities of the general GFAP candidates
To summarize, we managed to characterize the shRNA targets and sequence the PLKO1-GFAP candidate vectors (Table 1 and fig. 3). We found candidates that were able to knock-down GFAP in general, namely candidates 90603, 90604 and 90606 (Mouse) and candidate D3 (Human). Two
human candidate constructs are GFAP-alpha specific, namely D4 and C12. Virus particles were produced from the candidate vectors. A second generation protocol was used with third generation lentivectors and resulted in high enough titers to work with in future experiments.
To test the isoform specific knockdown of GFAP-delta and alpha we transfected U343 (human astrocytomas) or DBT (mouse astrocytomas) cells with the acquired candidate vectors. mRNA levels were checked to determine if transfection was successful. However, the mRNA levels itself are not enough to determine whether a candidate vector indeed causes a knock-down of an isoform, but it allows for an early assessment of the knockdown capabilities of the candidate vector. The amount of GFAP protein should therefore be assessed to quantify the actual amount of GFAP-alpha and delta protein that is produced in the different conditions that are tested. However this will be done in future experiments using lentiviral transduction to determine what the knock-down capabilities of these vectors are.
Validating the GFAP knock-down
The Q-PCR results for the standard lipofectamine tranfections deliberately have no error bars. Since a problem with lipofectamine transfections is that only about 40% to 50% of all the cells get
transfected. This results in large variability when using Q-PCR to quantify the amount of RNA. The cells that are not transfected keep producing mRNA, which is then detected with Q-PCR. Therefore, the amounts of GFAP-RNA measured between batches can vary greatly, depending on the
transfection efficiency.
A way to encounter this is switch to lentiviral transduction. Lentivirus has a far more efficient transduction rate than lipid transfections, since virus is actively transducing cells compared to the passive way of compounds like lipofectamine and PEI. This results in lentivirus having a transduction success of 90% compared to the 40% of a standard lipofectamine transfection, which will assure that noise is reduced to acceptable levels.
Virus production and titers
Virus was produced using a second generation protocol. However a third generation protocol was also tested. Regrettably no virus was produced using this protocol. Possible explanations are that having to transfect four vectors instead of three decreases the chance of all vectors being transfected into one cell. This would also explain why cells that were transfected with the viral vectors and a vector containing turboGFP showed green fluorescence, but failed to result in any virus after virus isolation. This means that the vector containing GFP was transfected into the cell but missing one of the virus vectors, which prevents the cell from producing any virus.
Except for PLKO1-tGFP the actual candidates containing shRNA oligo’s are not fluorescent. To still be able to titer these and assess the amount of transducable units (TU) a p24 ELISA is used. However a problem with p24 ELISA is that it will overestimate the amount of TU. This type of ELISA is sensitive to a protein that is part of the envelope of the virus. However, it does not reveal whether the virus particle contains the shRNA vector or if it is empty. To determine the actual amount of virus that is capable of transducing cells with the candidate vectors, the tGFP titer is used. This titer will give an indication of how many cells are actually transduced. This method tends to underestimate the amount of TU, because not all viruses containing the tGFP will transduce a cell. Therefore the mean
of both the p24 ELISA and the tGFP titer, which is determined by a cell count of fluorescent cells,is taken to determine the actual amount of viral units capable of transduction.
Future prospects
The acquired candidates for knocking-down GFAP in general and isoform specific will be used in future experiments to investigate different properties of (primary) astrocytes, while varying GFAP expression. Experiments that will induce mechanical stress will test how differences in amounts of GFAP affect the cells properties to deal with this stress. Other experiments involve life cell imaging experiments, which are designed to investigate the role of GFAP in the migration of astrocytes. Especially the effect of isoform-specific knock-down of either GFAP-alpha or delta on migration capabilities in neurogenic astrocytes are of great interest. All the experiments described are in vitro, although in vivo experiments are also planned. Enough virus was produced to facilitate in vitro experiments, however titers were not high enough to be utilizable in ‘in vivo’ research. However, when higher titers of the knock-down vectors are produced they will be used in vivo to knock-down GFAP (in general or isoform-specific) in the subventricular zone. These experiments can shed further light on how changes in GFAP expression affect neurogenic astrocytes and what consequences this entails for their function in vivo. As of yet no efforts in producing GFAP-delta clones resulted in construct capable of knocking-down GFAP-delta. However, some progress has been made by switching to the more stable PLKO1 vector (see ‘Materials and Methods’ section). Most of the cloning problems have been resolved and it is only a matter of time before GFAP-delta specific candidates are produced.
Table 1. Characterization of the general GFAP knock-down targets
Name Target sequence Target
exon GFAP unspecific Alpha specific Virus candidate 90603 (m) GAGAGAGATTCGCACTCAATA 4 No **
90604 (m) TCGTGTGGATTTGGAGAGAAA 3 No ** 90605 (m) GAGTGGTATCGGTCTAAGTTT 4 No 90606 (m) CTTTGCTAGCTACATCGAGAA 1 No ** 90607 (m) GATCTACTCAACGTTAAGCTA 6 No D1 (h) CAAGAGGAACATCGTGGTGAA 8 * Yes D2 (h) CCGCACGCAGTATGAGGCAAT 4 No D3 (h) GCCTATAGACAGGAAGCAGAT 2 No ** D4 (h) AGAGGTCATTAAGGAGTCCAA 9 * Yes ** C12 (h) CCCTTCTTACTCACACACAAA 9 Yes **
Table 2. Virus titers
Name Average TU/ul
Lv-cherry-ICTR 4.30*1010
PLKO1-tGFP (4/6/11) 5.83*109
PLKO1-tGFP (21/7/11) 6.94*108
12
Table 1. Characterization of the general GFAP knock-down targets. ‘Target sequence’ shows the specific sequence of the GFAP gene that is targeted by the shRNA. ‘Target exon’ indicates which exon is targeted. The next column tells which contructs are also targeting other genes. ‘Alpha specific’ shows which vectors target only specific for GFAP-alpha. ‘Virus production’ indicates what vectors are used for virus production after assessing their knock-down capabilities. (h)=human, (m)= mouse.
* See Fig. 8
** See Fig. 4, 5 and 6
Table 2. Virus titers. The first column describes the different vectors that were used to produce virus. Lv-cherry-ICTR is a construct with a non-targeting control shRNA (ICTR). The PLKO1-tGFP vectors are lentiviral constructs with a green fluorescent protein gene and used as a titering tool. The average amount of transducable units (TU) per microliter (ul) is shown in the second column. Titers larger than 109 can be
used for in vivo experiments. All virus was produced with the second generation virus protocol described in the ‘Materials and Methods; virus Production’ section.
References
Bovetti S., Gribaudo S, Puche A.C., De Marchis S. and Fasolo A. From progenitors to integrated neurons: Role of neurotransmitters in adult olfactory neurogenesis. Journal of Chemical
Neuroanatomy (2011) xxx–xxx.
Chakiath C.S. and EspositoD. Improved recombinational stability of lentiviral expression vectors using reduced-genome Escherichia coli. BioTechniques, (2007), 43:466-470.
Dirks P.B., Patel K., Hubbard S.L, Ackerley L., Hamel P.A. and James T Rutka. Retinoic acid and the cyclin dependent kinase inhibitors synergistically alter proliferation and morphology of U343 astrocytoma cells. Oncogene, (1997) 15, 2037-2048.
De las Mercedes Segura M., Kamen A. and Garnier A. Downstream processing of oncoretroviral and lentiviral gene therapy vectors. Biotechnology Advances, (2006)24;321– 337.
Manjunath N., Wu H., Subramanya S., Shankar P. Lentiviral delivery of short hairpin RNAs. Advanced
Drug Delivery Reviews 61, (2009) 732–745.
Middeldorp J. and Hol E.M. GFAP in health and disease. Progress in Neurobiology, (2011) xxx–xxx. Middeldorp J., Boer K., Sluijs J.A., De Filippis L., Encha-Razavi F., L. Vescovi, Swaab D.F., Aronica E., and Hol E.M. GFAPδ in radial glia and subventricular zone progenitors in the developing human cortex.
Development 137, (2010), 313-321.
Platel J.C., Gordon V., Heintz T., and Bordey A. GFAP-GFP Neural Progenitors Are Antigenically Homogeneous and Anchored in Their Enclosed Mosaic Niche. Glia, (2009), 57:66–78.
Perng M., Wen S., Gibbon T., Middeldorp T., Sluijs J., Hol E.M. and Quinlan R.A. Glial Fibrillary Acidic Protein Filaments Can Tolerate the Incorporation of Assembly-compromised GFAP-δ, but with
Consequences for Filament Organization and αβ-Crystallin Association. Molecular Biology of the Cell,
(2008) Vol. 19, 4521–4533.
Roelofs R.F., Fischer D.F., Houtman S.H., Sluijs J.A., van Haren W., Van Leeuwen F.W., and Hol E.M. Adult Human Subventricular, Subgranular, and Subpial Zones Contain Astrocytes With a Specialized Intermediate Filament Cytoskeleton. Glia, (2005), 52:289–300.
Sexton M., Woodruff G., Cudaback E., Kreitzer F.R., Xu C., Lin Y.H., Möller T., Bai M., Manning H.C., Bornhop D. and Stella N. Binding of NIR-conPK and NIR-6T to Astrocytomas and Microglial Cells: Evidence for a Protein Related to TSPO. PLoS ONE (2009) Vol 4; 12, 1-9.
